Elsevier

Journal of Catalysis

Volume 227, Issue 1, 1 October 2004, Pages 217-226
Journal of Catalysis

Promoting and poisoning effects of Na and Cl coadsorption on CO oxidation over MgO-supported Au nanoparticles

https://doi.org/10.1016/j.jcat.2004.07.009Get rights and content

Abstract

Density-functional theory has been used to study the CO oxidation reaction at model Au/MgO catalysts modified by either Na (an electron donor) or Cl (an electron acceptor) dopants. In agreement with experimental observations, Cl is found to act as a poison, making both the adsorption of O2 and the formation of the CO ⋅ O2 intermediate complexes that mediate the formation of CO2 on these model catalysts more difficult. The poisoning effect of Cl has a long-ranged character, reaching at least two Au sites away from the adsorbed Cl atom. On the other hand, Na is found to be a promoter, enhancing both O2 binding and CO ⋅ O2 formation. Its promotive character is, however, local, involving the formation of strong Nasingle bondO bonds.

Introduction

While the surfaces of bulk gold are known to be catalytically inactive [1], oxide-supported Au nanoparticles display remarkable catalytic properties for several interesting reactions [2], [3], with low-temperature CO oxidation being one of the most extensively studied examples [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16]. The crucial step in catalytic oxidation is O2 adsorption, and the inertness of gold bulk surfaces is understood from the poor affinity toward molecular oxygen [17], [18], [19], [20], [21]. Much effort has been dedicated to understand the origin of the unusual reactivity of nanosized Au particles. The influence of the supported catalysts' properties such as size, type of support, and method of preparation has been explored. As a result, it is now generally accepted that a small particle-size distribution (<5nm) is necessary to obtain active catalysts. Also, sizeable differences between active (e.g., TiO2 or Fe2O3) and nonactive (e.g., MgO and Al2O3) oxide supports have been measured [9], with a superior ability to attract and activate molecular oxygen for the former ones. Recently, theoretical studies have indicated that different reaction mechanisms are dominating in the two cases [13], [14], [15], [16]. Due to the weak O2 binding, it has been suggested that the reaction mechanism for CO oxidation involves the formation of intermediate CO ⋅ O2 complexes, prior to the oxidation of the CO molecule and CO2 formation [12], [13], [22]. On a nonactive support material, an Eley–Rideal reaction mechanism for CO oxidation at the Au/MgO perimeter has been proposed, showing low reaction barriers for both CO ⋅ O2 and the subsequent CO2 formation [13], [14]. In contrast, studies of this reaction at the Au/TiO2 perimeter reveal that a Langmuir–Hinshelwood mechanism is possible, mainly because of the increased ability of the support to attract and activate oxygen molecules [15], [16]. However, common for both cases is the active character of low-coordinated Au atoms at the Au/support perimeter interface.

For the supported Au systems, one rather unexplored topic is the influence of coadsorbates on the catalysts' performance. The activity for CO oxidation has been shown to depend crucially on the catalyst preparation. In particular, the activity is sensitive to traces of chlorine, left from the catalyst synthesis when using, e.g., HAuCl4 as precursor. The poisonous effect of Cl is believed to be twofold [23]: First, during the calcination of the prepared catalyst, Cl enhances sintering of Au particles. Secondly, Cl hinders reactant adsorption by site blocking and/or by long-ranged electrostatic interactions. Furthermore, electronic effects could be important, where Cl drain electrons from the Au particle. The site-blocking scenario is supported by measurements of an activity suppression for the CO oxidation reaction at a Cl/Au molar ratio even as low as 0.0006 [23]. The electronic effect modifies the activation of adsorbed reactants, and especially the activation of the adsorbed O2 molecules, where the formation of a superoxo or a peroxo ion is crucial for the continuous reaction. In contrast to Cl, other coadsorbates could act as promoters. For example, Häkkinen et al. [6] have recently studied the effect of adding an electron donor (Sr) to small MgO-supported Au clusters. It was found that the addition of Sr significantly increases the amount of CO2 produced per deposited cluster. This was explained by enhanced adsorption and activation of O2 on the Sr-doped systems.

In this paper, we use a first-principle technique to study the effect of adding either Cl (a one-electron acceptor) or Na (a one-electron donor) to a model Au/MgO catalyst. By using two different catalyst modifiers, we are able to study both poisoning and promoting effects within the same computational scheme. Several aspects regarding the catalytically activated CO oxidation reaction will be considered, including adsorption of Cl/Na on the Au/MgO model system, coadsorption of CO and O2 together with Cl/Na, and CO + O2 reaction in the presence of Cl/Na. The selection of MgO as support is based on two considerations: First, the structural simplicity of MgO reduces the size of the model systems, allowing for an extensive study. Second, the effects due to the poisoning or promotion of the Au nanoparticles by themselves are expected to be largest on a nonactive material as MgO.

Section snippets

Computational method

All calculations are performed using the density-functional theory (DFT) in the implementation with plane waves and ultrasoft pseudopotentials [24]. The results are obtained in the GGA-RPBE density-functional formalism [25]. RPBE is known to improve the description of molecules and the adsorption of molecules on surfaces over other GGA formulations [25]. More specifically, we believe that the binding of oxygen to gold is better described with this functional, as it reduces the overbinding found

Model systems: Au/MgO, Na/Au/MgO, and Cl/Au/MgO

Different aspects of modeling Au/MgO interface boundaries have been discussed in Refs. [13], [14]. In this study, the main focus is on the CO oxidation reaction over a modified Au/MgO model catalyst. The aim is to explore the effect of coadsorbates on the stabilities of the involved species and activation energy for the CO + O2 reaction. As mentioned in Section 1, both a single electron donor (Na) and a single electron acceptor (Cl) are used as catalyst modifier. The model catalyst systems are

Adsorption of reactants

In this section, the relevant steps preceding the actual CO oxidation reaction are discussed. The main focus is on relative stabilities among the reacting adsorbates, CO and O2, adsorbed at the modified Au/MgO edge.

CO oxidation reaction

Having established the adsorption energetics for the involved reactants, we now continue with the actual CO oxidation reaction. To close the catalytic cycle, the important following oxygen abstraction reaction step will also be considered subsequently.

Conclusion

The effect of coadsorbates (Na and Cl) on the CO oxidation reaction over a Au/MgO model catalyst has been studied using density-functional theory, investigating both promoting and poisonous effects. The interaction between Na and Cl with the Au/MgO catalyst is found to be strong. Charge transfer is observed between the Au metal and the coadsorbates, indicative of an ionic binding forming a Naδ+ and a Clδ upon adsorption. The formation of charged adatoms prevents clustering at the edge. For Na,

Acknowledgments

NORFA is gratefully acknowledged for financial support (P.B.) (030449). So is the Danish Research Councils and Dansk Center for Scientific Computing (L.M.M. and B.H.).

The Competence Centre for Catalysis is hosted by Chalmers University of Technology and financially supported by the Swedish Energy Agency and the member companies AB Volvo, Saab Automobile Powertrain AB, Johnson Matthey CSD, Perstorp AB, Akzo Nobel Catalyst, AVL-MTC AB, and the Swedish Space Corporation.

References (43)

  • M.M. Schubert et al.

    J. Catal.

    (2001)
  • A. Wolf et al.

    Appl. Catal. A

    (2002)
  • D.A. Outka et al.

    Surf. Sci.

    (1987)
  • J.M. Gottfried et al.

    Surf. Sci.

    (2002)
  • H.S. Oh et al.

    J. Catal.

    (2002)
  • J.V. Barth et al.

    Surf. Sci.

    (1995)
  • P.S. Bagus et al.

    Surf. Sci.

    (1992)
  • B. Hammer et al.

    Nature

    (1995)
  • M. Haruta et al.

    Chem. Lett.

    (1987)
  • M. Haruta

    J. Catal.

    (1997)
  • M. Valden et al.

    Science

    (1998)
  • A. Sanchez et al.

    J. Phys. Chem. A

    (1999)
  • H. Häkkinen et al.

    Angew. Chem. Int. Edit.

    (2003)
  • J. Hagen et al.

    Phys. Chem. Chem. Phys.

    (2002)
  • W.T. Wallace et al.

    J. Am. Chem. Soc.

    (2002)
  • N. López et al.

    J. Am. Chem. Soc.

    (2002)
  • Z.-P. Liu et al.

    J. Am. Chem. Soc.

    (2002)
  • L.M. Molina et al.

    Phys. Rev. Lett.

    (2003)
  • L.M. Molina et al.

    Phys. Rev. B

    (2004)
  • Z.-P. Liu et al.

    Phys. Rev. Lett.

    (2003)
  • L.M. Molina et al.

    J. Chem. Phys.

    (2004)
  • Cited by (59)

    • Controlling the metal work function through atomic-scale surface engineering

      2022, Applied Surface Science
      Citation Excerpt :

      Gross and co-workers used periodic DFT calculations to elucidate coverage effects of halides adsorbed on Pt(111) and Cu(111) and how these adsorbates induce a dipole change at the surface interface [62,63]. Broqvist et al. studied the CO oxidation reaction over model Au/MgO catalysts modified by either Na or Cl to investigate dipole effects promoted by dopants on supported catalysts [64]. Chou and co-workers [65] developed a general model, namely the orbital-overlap model, to describe their observed coverage dependence in computed work function for various Cs-covered transition metal surfaces, including Ni.

    • AuCu alloy nanoparticles supported on SiO<inf>2</inf>: Impact of redox pretreatments in the catalyst performance in CO oxidation

      2017, Catalysis Today
      Citation Excerpt :

      On the other hand, after the reducing treatment (RED sample) no traces of chlorine could be detected on the sample. Considering the well-known poisoning effect of chlorine in CO oxidation over Au catalysts [32,38] we can speculate that the oxidative treatment resulted indeed in the removal of organic surfactants, but it was not effective for the removal of residual chlorine from the surface of the Au0.75Cu0.25 NPs. On the opposite, the high temperature treatment in a reducing atmosphere effectively removed the chlorine from the Au0.75Cu0.25 NPs, thus enhancing significantly the catalytic performance and leading to a stable activity.

    • Catalytic abatement of CO over highly stable Pt supported on Ta <inf>2</inf>O<inf>5</inf> nanotubes

      2014, Catalysis Communications
      Citation Excerpt :

      Indeed, some works stressed on the poisoning effect of chlorine contamination in CO oxidation reaction. Cl was found to act as a poison, making both the adsorption of O2 and the formation of the CO–O2 intermediate complexes more difficult [26–28]. Thereby, the crystalline Ta2O5 NTs seems to be a more promising support for Pt catalyst since this material demonstrated higher activity and no contaminations on the surface.

    View all citing articles on Scopus
    View full text